Abstract:
At present, text sentiment analysis models rarely incorporate personality cues. However, users with different personalities have different emotional expressions. Based on the Big-Five personality model in psychology, this paper proposes a Weibo sentiment classification model PBiLSTM that combines personality cues.The model integrates the sentiment features of Weibo sentence text with the user's personality cues, thereby adding a new dimension of sentiment classification. At the same time, it uses BiLSTM to extract the advantages of the global features of the text. This method effectively improves the model's ability to classify emotions. The experimental results show that the accuracy of PBiLSTM method can reach 93.68%, and has achieved good results on multiple performance indicators.